Artificial Intelligence - Transformative Tech or Management Fluff?

Today, we have a special guest, Alex Alexander, to discuss Gen AI. Alex runs a tech start-up, Xoots and has a glittering career in technology across Europe, working for brands including Emirates and Net A Porter. He's advised companies and governments on all things AI and we are lucky to have him join us to explore this hot topic.
 
 He will help educate you on what Gen AI is and how it works, explore some of its uses and potential, and finally, Alex will share his insights on how you can get started and exploit the power of Gen AI.
 
 If you've wondered what it is or how to start on the journey, you mustn't miss the opportunity to hear from a genuine AI Guru!
 
 If you would like to know more about Gen AI or get help from Alex, you can get in touch via his company, Xoots or by Linkedin
. 
 XOOTS – Your digital transformation co-pilot

Alex Alexander | LinkedIn

Transcript

James: 0:03

hello, I'm James.

Jimmy: 0:05

Hi, I'm Jimmy.

James: 0:06

Welcome to A Job Done Well,

Jimmy: 0:08

the podcast about the world of work and how to improve the daily grind

0:12

Hi, we've got a couple of interviews coming up, including this one that we've recorded remotely and the sound quality isn't quite up to our usual standards. We will continue to test and learn and find ways to improve the sound quality of interviews, but we thought the content and insights these interviews have are excellent. And we don't think it will really spoil your enjoyment. So we want it to still share them with you, but we thought it was worth mentioning upfront. She don't spend the first five minutes thinking there's something wrong with your headphones anyway, on with the show.

Jimmy and James (2): 0:45

Right. Good morning. How are you doing? I'm doing well. How are you doing, James? Very good. Thank you very much. Got something slightly different for you today. Yeah. Yeah, we are very, very lucky to have an expert. And when I say expert, I actually, we call ourselves experts and we call ourselves gurus. And that's us overstating things, but we genuinely have an expert and a guru in artificial intelligence with us today. So we are very lucky to welcome Alex Alexander. Hi, Alex.

Alex: 1:13

Hello, I'm flattered and first of all, thank you for having me and as a podcast co host for Beyond the Tech podcast, it feels really strange for me to be a guest today and not ask any questions. So if I misbehave and try to ask questions, then you please stop me.

Jimmy and James (2): 1:31

effectively. We've got three podcast hosts what could possibly go wrong, but Alex we know a bit about your background, your expertise. You advise, companies, governments, all sorts on this subject. So, just for the benefit of the listeners, do you want to give us a, a quick. rundown of your experience. But just to give a flavor of what we're going to cover off today, we're going to focus a little bit on education. What is, what do we mean when we talk about AI and Gen AI in particular? We're going to explore the things that work and things that don't work. And you're going to also help us with some advice on how you can get going and how you can unlock the power of AI. But first, for our listeners. Tell them, why should we be listening to you about this? What's your background?

Alex: 2:19

Yeah, fantastic. so first of all, I've been in the tech space, for too long, perhaps over two decades now, and it's been quite a journey. Um, I've had the pleasure of leading technology teams for some of the, world's most renowned companies in UK. I worked for likes of Sainsbury's and Barclays bank. Uh, I worked for Walmart in the U S uh, yukes, uh, Net A Porter. They've, online fashion retailer in Italy. Group in Dubai and at the venture group, the global marketplace, in Spain and a few others. Uh, and I've been very lucky that these roles have taken me to seven countries across three continents, which has been an amazing, experience. But most importantly. My rules have always been about creating something new, often a new digital product across, you know, industries such as consumer retail or retail banking, or even aviation. Uh, and actually it was this passion for creating something new that led me to set up my own generative AI tech startup called Zoots, uh, last summer, but I'm not new to AI. Um, I've got over a decade of experience, in the field of AI, integrating AI into customer interactions, enabling personalizations, but also integrating into, uh, operations and supply chain and improving, operational efficiency. So really glad to be here today.

Jimmy and James (2): 3:42

Yeah, thank you, Alex. We're very lucky to have you. So just to start off with, I know this is really basic stuff for you, but some of our audience will not have come across, Gen AI particularly, or there are lots of people are pretending they understand it. It's the, it's a great word. It is buzzword du jour. And, uh, yeah, so people don't have to pretend to understand it. what do you mean by a gen AI and how does it work?

Alex: 4:11

Absolutely. I think this is true because there's a lot of, buzzwords around, but, uh, I would say generative AI is pretty fascinating and, uh, because it's made up of this massive, big data driven models, that, are behind the generative AI technology as we know it, and they're capable of creating content. And when I say content, I mean a wide range of content from images to videos to text to music. It's like having a super creative, um, that can generate all sorts of things. and, um, it excels at creating this diverse, content across, many industries. So think about it as, art industry, content creation, media industry, music industry, drug discovery, uh, software development, which is obviously one of the areas that I've, worked on as well. So many, many industries. And often when we talk about generative AI, we think about tools such as OpenAI's ChatGPT or Google's BARD are perfect example of the power of generative AI because, they generate human like texts, which opens up, you know, possibilities for creating content, improving customer service and, a lot more. So if I had to sum it up, I would say generative AI is. Is a transformative technology and is changing the way that we create. And also interact with content and it's really excited to see, you know, where it's going to head in terms of, evolution of it, uh, in the next few months and few years.

Jimmy and James (2): 5:46

So the risk of saying something really stupid, Alex, but so generative AI is generating stuff. As opposed to AI, which is not necessarily generating stuff.

Alex: 6:00

Well, there is a distinct difference between, you know, traditional AI, I'll call it, versus generative AI. And, um, the way I would, describe it, and by the way, James, it's a great question because often people get, confused is that traditional AI uses machine learning algorithms to analyze data, uh, for tasks like, market predictions and, and that could be for, financial market predictions that could be for healthcare, could be for manufacturing, and for things like fraud detections and medical diagnosis. And machine learning, uh, which is a subset of AI or often referred to as ML, uh, so shortened for machine learning, is probably the most common form of AI we see across various industries. So when you see product recommendations on an e commerce site that are generated through machine learning, if you're seeing, um, social media feeds being curated, that's machine learning. Uh, so it's great for, you know, all kinds of examples that I mentioned, but also including, um, fraud detection and supply chain management. On the other hand, generative AI, as you rightly mentioned, James, is about, uh, creating content. It generates content, and as I said, content means images, videos, text, and even music. so its application obviously touches many industry from content creation, media, music, So, uh, I think that's it. Uh, drug discovery, software development and many others, but a lot of focus, and headlines, I would say has been on the subset of AI called large language models, uh, or generative AI, uh, and these generative AI tools, like chat, GBT, uh, when, When we refer to AI, often people think that's AI, but actually it is really important to remember that traditional AI has been around for over 50 years and doing some amazing work. Uh, so we should not forget that AI is a new, generative AI is new. and one interesting, KPI to just put it into perspective and understand the scale of it. Chat GPT 4, which is one of these so called large language models has about 1 trillion parameters that shows the scale and the size of the, power of the data that he has, and a human brain has around 100 trillion parameters. So perhaps it's not long before these large language models are catching up with humans, but it's not quite, that sequential because humans can do and continue to do things that, AI cannot do. So in a nutshell, All of these technologies fall under AI umbrella, but they have different focuses and different applications.

Jimmy and James (2): 9:01

So just, sorry, because I'm intrigued, the history lesson. So you say it's been around for 50 years. So can you give us some examples of things which, you know, which maybe we grew up with that we don't really think of as being AI, but in actual fact were. So, what was happening in the 90s, what was happening in the 2000s.

Alex: 9:17

Well, I mean, we've been all familiar with, Amazon website or other websites that we go to that, when we go for the second and third time, it recognizes us. So that's personalization, and other examples of, You know, using AI, but also even, you know, some elements of the generative AI. When we do Google translate that is actually using generative AI, but of course, Google translator has been around for many years, but it was the early form of generative AI. So all of these technologies in some shape or form have been around. Another example of it, dynamic pricing, is using a generative AI. Um, so, but traditional AI is about algorithms that predict where generative AI is about content creation.

Jimmy and James (2): 10:02

Yeah. And I think so our experience with AI would have been in prediction. So we both used to work in insurance, but making a prediction on whether or not something is fraudulent, for example.

Alex: 10:12

Exactly. Fraud detection is a key example of traditional AI that we have seen for many years.

Jimmy and James (2): 10:18

So one of the things that gets me is people listen to all the hype about AI and And think, well, it's a machine. Therefore, it must be right in actual fact. It's a prediction. It's not a statement of fact. A lot of this is about predictions. Is that I got it right?

Alex: 10:37

Yeah, absolutely. Right. I mean, it very much depends on the algorithm very much depends on the data that it's used, uh, to build this prediction. Absolutely. At the end of the day. A human has trained these models and written the predictions. Yeah, so you're absolutely correct. that, uh, it very much depends on the algorithms and the data that it's, consumed.

Jimmy and James (2): 11:02

I is going to take over the world. It's the end of civilization. I guess what is the potential of AI?

Alex: 11:09

Well, in my view, AI is very transformative. But many people say that AI is going to replace all jobs and, and take away all the skills that we have. I'm an AI practitioner, which is not of that view. My view is that AI will enhance and augment our capabilities, our jobs, hence our productivity. So I'm a great believer of augmentation. So I'm a believer of augmentation. So human intelligence combined with AI to create this exponential transformation capability that, this new technology is capable of providing, but it's not about replacing, it's about augmenting and enhancing.

Jimmy and James (2): 11:57

And I think that's really interesting. So when I have used AI tools, so for example, on our website. For this podcast, you will see there were lots of images and they're all generative AI. They all come from a system called mid journey, but it's about the interaction between the human and the AI, I think, which is what gets you the best outcome rather than just expecting the AI to give you the best thing. It's a bit like, uh, now I'm going to be really nerdy, but computer chess. A computer will always beat a man at chess now, but a human computer partnership is invariably a much more forbiddable chess playing machine. So it's about the combination between the human and the machine, I think, which is where the real power is.

Alex: 12:40

Absolutely. I totally agree with you, James, on that one. And the key to success in the future of, you know, skills, jobs, is about, using AI as an enhancer, as augmentation. And it's not, it's not about competing with AI. It's about working with AI to, to enhance what we do and become better in our jobs and become far more productive.

Jimmy: 13:05

The show you're listening to right now brings you insights to help you improve performance and enjoyment of work.

James: 13:12

It takes a long time to produce, but we think it's well worth it. It's all in the name of helping you getting the job. done well.

Jimmy: 13:19

All we ask in return is that you share the link to this show with someone that you think would benefit from it. And if you haven't already, Click on follow wherever you're listening to this podcast right now. Anyway, let's get back to the show.

Jimmy and James (2): 13:35

But just to, get in ahead of James saying is to, to quote public enemy from back in the day, don't believe the hype, we've heard, machine learning, robotics, big data, is this just the next big thing from the tech world?

Alex: 13:50

great question. and, I have also seen just like you a lot of hype over the years around various technologies. Uh, and many of them never materialize or some of them took a long time to, you know, become reality. But generative AI, it feels different. I mean, just to put it into context, it took chat GPT, you know, when it was launched in November, 2022. And I remember that time. Precisely because I was curious and I was not convinced, but it hit 1 million users within the first five days. And it took Netflix, for example, two and a half years to reach the same milestone. And it's this kind of rapid adoption, but also a rapid change of this technology, because it's not just that version of chat GPT that we are using now. There's been many, many other iterations of that since then. And it is. transformative because, it makes us more productive. And it's not just in, for our personal use, but also for professional work, for industries such as healthcare, finance, journalism, and education. So I would say, it is different. But, but we also have to be cautious about using it because it's not easy to adopt it properly because it requires, good data management. And it also needs to make sure that we are aware of the ethical use of AI as well.

Jimmy and James (2): 15:16

Let's come, let's come back to the adoption in a, in a moment, but can you give us some good examples, some use cases of where James AI is working on and has. Deliver transformative change and some of the stuff that you see in some of the stuff that you're doing at zoos and maybe some examples of where it hasn't done quite as good a job.

Alex: 15:38

Yeah, great. So let me sort of share with you, a case study from my own experience with Zoots and one of our products, Talent X, because we built an AI powered talent interviewing platform, and we believe this platform will transform the talent selection process because we believe it's going to make it, you know, a thousand times more efficient, but also less biased and more transparent, but also at the same time, creating a more enjoyable experience for the candidates as well. so we use AI to effectively do interviewing. So it creates efficiency, accuracy, and, and we've seen, the outcome of that, in terms of not only speed, but also quality of the candidates and that position that it gets in terms of finding the. The right talent that you're looking for, but probably a second example that I can mention again is to do with Zoots is actually our team. Because, as I mentioned that we set up the startup less than a year ago, so last summer, but we have a team of hundred employees, but only six of us are humans. And the other 94 is what we call AI agents. And they handle various tasks for us from, taking meeting notes to booking appointments with client appointments, doing research for us, and doing all of these things so that the six humans. are a lot more productive in terms of what we do. In fact, some of the software development that we have done, we've done it at half the time, it would have taken me pre generative AI. so all the estimates that we did at the beginning of this journey is actually in practice is, is half just because, we're using, human and AI together to augment the work that, uh, we are doing. In terms of other examples, AI,

Jimmy and James (2): 17:30

just to pause one second just on on your talent X example only because I've seen it. you took me through it. I mean, I always hate interviewing. I don't have the patience for it. I, I decide within. 30 seconds whether I like someone or not. and so I, I was intrigued by your platform for two reasons. One, because it does a job I don't like doing. I don't think I'm very good at, but also it was really clever in that you had identified the key skills and the answers and the things that people should say to get a really good answer. So it would say I'm hiring for a complaints handler. Here's the key things I need for a complaints handler. The person's interviewed by the AI. via a video chat and then it dispassionately and objectively evaluates their interview rather than me, after 30 seconds, I like decide that that person is boring and I want to carry on. So I think it's, it will definitely in that example, um, be much more efficient and give a better quality outcome. Can I say, I haven't seen it, so I'm talking from the position of ignorance, that doesn't normally stop me. Surely then, Alex, with something like that, it's only as good as the information that is put into it. So, is it really interesting you say it removes bias, is it not as biased as the person who has given it the information from the get go, and what a good candidate looks like, or is there a way around that?

Alex: 18:59

well, I think this is, I mean, it's a great question. So let me just try to break it, uh, break down the answer. But first of all, it's back to human and AI working together.

Jimmy and James (2): 19:10

Yeah.

Alex: 19:10

So humans saying, what is the best practice type questions and best practice answer for a certain question, because that's not bias. It's just. This is, based on years of experience that, you know, whether it is, somebody, a practitioner in technology or operations or, compliance, whatever it may be. So that's just, um, understanding what is, good questions and how do we, test the knowledge AI uses is knowledge of that, information. to analyze the answers that a candidate gives verbally. And that's another important part point as well, because the candidates can answer the questions in one of 65 languages if they want to. But again, to reduce the bias of accents, languages, AI will only, evaluate speech to text, either translated, voice into English language. So it doesn't create a bias towards the languages or accents. So It's the power of human and AI working together to say, what is a good answer? And AI being able to analyze that and then transcribe that in terms of, this was the question, this was the answer that I was given, and this is my analysis of whether the answer fit the mark

Jimmy and James (2): 20:20

like the story about women playing in orchestras. So when, um, when they were hiring for orchestras, they used to do performances. And the number of women who got through to play in the orchestras was actually quite low. What they then moved to was a position where you did the performance, but you played behind the curtain so people couldn't see it. And then when they played in that scenario, many more women went to orchestras because by putting the curtain in front, you took the bias out. People couldn't see it was a man playing. Sorry, that sounds a bit random, but in a lot of ways, that's very similar to what you're saying because you put in the criteria, but Yeah, you could put in by this criteria, but you're less likely to do that.

Alex: 21:02

Exactly. And, and I mean your analogy or say your example is, uh, very relevant because another thing that we do to eliminate or reduce that bias is that, you know, everything with a candidate starts with a cv. And of course, AI will want to see the cv, but it will, we will only allow AI to see the CV after it analyzed the answers after the interview. because we don't want it to create a bias towards the company that a person had worked in, uh, the location that they live in, or, the languages they speak or they don't speak. So as a result, there is a sequence by which we can do things, which actually is the opposite of how humans do things. you know, perform that activity to reduce the bias, and create transparency.

Jimmy and James (2): 21:49

I think I think the example is a good one because I think one of your key points here Alex is it's the Link between the human and the AI and how they work in synergy that is important So in your example, I like interviewing that doesn't mean that I don't have to have any input into this That the AI is just going to find me Great customer service agents. It's not going to do that. I've got to teach it and work with it on what good looks like first and then and then once it's got that, it can then churn through lots of interviews without bias. And give me objective outcomes, but it's not, it doesn't naturally know what a good customer service agent looks like, in my environment, in my context,

Alex: 22:37

You're absolutely correct. You define it exactly and precisely because what an AI does can scale this, can scale what that, what you described in terms of processing Thousands of candidates a minute to do that at a scale at a level of efficiency that humans cannot do and produce the most consistent set of transcripts of interview notes that humans are not able to produce because it's not. It's just very objective just because it's processed, the answers based on what you just described to come up with, what good looks like. So that's exactly it. And we will, we have seen in the last, two and a half months that, we've used it live. And of saves in excess of 3, 000. per hire, but it's not just the cost savings is actually the better experience, you know, the more transparency, and so on. That is also great as well. So faster time, a thousand times faster than humans, but, also, better experience.

Jimmy and James (2): 23:41

and that's, that's when I mean, that's a great story, but I mean, that's when I AI works well, but the part of the trouble is, at the moment, everyone's getting involved in AI and everyone's got every organization's got their AI projects, and one of my pet hates at the moment is chatbots. That are just seem to be there to defeat you and stop you from getting answers and stop you from getting through to the people, what's going wrong with with those? Why are they? Why are they so frustrating to work with?

Alex: 24:13

Yeah. I mean, just like you, I get frustrated with those, but I think it's just to do with poor training of those AI models. And in fact, um, Jenny, I. itself is not without its hiccups. Um, and you asked earlier about what's the examples of failures of Gen AI. A notable example that I saw back in January of this year was, uh, with the parcel delivery DPT. They had to disable part of their online support chat bot, after it started insulting the customers. so can you imagine that, for customer service. So the chat bot that was used. Uh, to answer questions, for the customers started behaving unexpectedly, uh, after updates that, DPD had done. Uh, so not only just didn't answer the questions, it started, swearing at customers and it also criticized the company. So, they, DPD took it down. So these are, you know, fundamental issues in not properly. training AI models because AI works with, in a repeatable way once he's trained based on a set of data and customer and sort of business's context for those customers. So if those things are not in place, accidents such as this, will, uh, will happen.

Jimmy and James (2): 25:26

The tool is really powerful. But you need to be very careful, because it will do exactly what you ask it to do. I mean, the story, I don't know, I think I've got this right, I think it was Microsoft, they um, Built a tweet bot and they gave this thing the challenge of getting as many followers as possible in a short time as it could. And it was remarkably successful. And all he did was started to tweet a load of Nazi homophobic stuff. Got lots and lots of followers, but it wasn't quite what Microsoft had planned. But he did exactly what they wanted it to do. So the, and it might as well be a chainsaw, you know, a very powerful thing. But he needs to be handled with care. That's really what you're saying to us.

Alex: 26:04

Exactly. Exactly. I mean, I'm going to use a term which may sound a bit, geeky, but I will explain it. So everybody refers to this notion of prompt engineering. And all that means is that you've got to guide the AI to say how to behave, provide the context, provide it with, you know, what are the things that he wanted to do in what way? So what is the tone of voice for the business? How customer interactions need to be, action. How do you deal with the angry customer based on what they put into the text? Uh, so Jimmy was sort of saying that you get frustrated that sometimes you may actually want to get an answer. And I think AI needs to. We train to say at some point you need to pass it to a human so that the customer does not get frustrated. But all of that is to do with that training that needs to be done. If they haven't done that, they haven't provided it with a tone of voice and how to deal with customers, then problems like this will happen. So you're, you're correct, James.

Jimmy and James (2): 26:59

Now, Alex, many of our listeners will either be on the AI journey or thinking about the AI journey. Everyone has to do that at the moment. So it's like James said, it's the buzzword du jour. Um, what would be the advice that you would recommend, you know, people follow in order to help set themselves up for success on this journey?

Alex: 27:24

Yeah. So, I mean, just to wrap up this particular topic, I mean, everybody needs to know that generative AI, it's a bit like a Swiss army knife. is good for a lot of things, but not necessarily great at everything. So we have to be careful and we have to make sure that we pick the right tool for the job. Um, so whether that job is, about, we want to differentiate ourselves from others. Well, you need to make sure that, you train your AI to be differentiating and that means high quality business data. So that's, tip number one. Tip number two is that, you need to know exactly what data, and how accurate that data is that you're feeding to AI because gen AI needs data. That's, how you can process. So you've got to be careful, super careful about risk of reputational damage. so that's super important. Of course, if you haven't got any data governance in your organization, you have to create it because without data governance, you know, wrong data could get into, get fed into Gen AI and that obviously could have legal precautions and reputational damage. So it's so important to know that. And the third thing is that generative AI models, as we know it, They're effectively, uh, generalists. So you need to make them specialists and specialists for what you want them to do for the company purpose that they have. So these big models are Generalists you really need to make sure that you train them, you fine tune them to do what you want them to do based on your company's data and your company's context. And, and I would say the fourth thing is that you've got to get. If there is something to do with customer, you need to have the holistic customer data ready. Because, if, your data, it's, split across, 50 systems. but, five of those, systems you don't have access to. Well, that's, that's a failure point because, AI needs to have all of those data points because, it doesn't know for those data points that you are not providing it. What. issues that, that a customer may have had that you, the AI will not be aware of. And the last thing is, AI is about, you know, creating a cross functional team. So you need to get your team, mobilized and you need to make sure that they are educated on AI, what it is good at, and also what it is not good at. So I think that's the important part.

Jimmy and James (2): 29:50

But then if I was to. So there were about five or six things you said there. If I was to summarize it down to two, just because I'm a man who can't hold too many thoughts. But the one is you really do need to get your company data in a place where it can look at it. And then the second one is, is the human interaction with that. But you need to be very skilled at being, well, it's being clear about what you want the AI to do. Because if you're not clear about what you want it to do, or you don't have the data, but you run a little bit behind into nothing, or you'll get a whole load of, um, unforeseen consequences. I think it's interesting, Alex, I mean, your advice is very, very clear and very compelling in terms of, how you go about doing it. I think it'd be fascinating to know how many of these companies that are already on the AI journey have actually followed some of your advice because in my experience, all the, all the organizations I work for or worked with, and it's been quite a lot, none of them have their data straight yet. I bet a lot of them are on the, AI journey. So, the risk being. on one end, you're, you're slightly suboptimal in how you implement it and the benefits you get from it. And then we end up with another, false dawn of this is how technology can change our lives. Or at the other end, you could make a complete mess of things and expose your company to some significant risk by going down this route. Yeah, it's interesting. There's Drifco. Hang on, see if I can remember it. There is nothing quite so useless as doing the wrong thing more efficiently. it's got the real power. Yeah, but unless you're really clear about what you want to be done, Your Honour, yeah, I've set the height into nothing.

Alex: 31:37

Yeah, I think, um, I think spot on, uh, because having the right data and actually asking AI to, um, operate within the context of your organization, what is it that you want it to do? But, I know you summarized it very well, James, in terms of five points that I mentioned into two points, which are actually correct, but actually I can narrow it down to one point

Jimmy and James (2): 31:57

Oh, go on. Yeah. Even better. I like it. That'll work for me now.

Alex: 32:00

to really make use of generative AI, you need to have data quality as a priority. Because generative AI without high quality data is just fluff. That's my one line saying, what is the main thing that everybody needs to know?

Jimmy and James (2): 32:17

which will be interesting because how many organizations have good, good, high quality? Well, no, I've seen 30 odd years of management fluff. I think it'll fit right in. So, yeah. So given that, so I've worked for six organizations, I think two or three of them might have the data in a place they could use it. So my sample is half organizations don't have the data. However, there will be enormous political pressure in all organizations to get something going with AI. So if you're in the position whereby you've been tasked with getting something going And the organization doesn't have the data because it's a big old job to sort that out What can organizations sensibly do that will prove that it's useful? Where would you where would you cut it? So there's the recruitment application we've been talking about and I can see that. Is there anything else you think organizations should do? Yeah,

Alex: 33:13

Yeah, I think it narrows down the experiments. Uh, I think you really need to know what data, however narrow they may be, that could help you. But of course, organizations that are in, medical care should avoid if they haven't got the data, they should avoid doing any experimentation because we're dealing with, patients data and patients records. But, you know, it is possible to experiment in a small scale, for the areas that have no ethical implications. They have no. Reputational damage implications. So you really need to narrow down your scope to, based on the data that you can reliably, say it is of high quality. So it just narrows down. So, it could be things such as, supply chain optimization, looking at, detection of, faulty. items. So avoid things which is as a director, you know, interactions with a customer, but more focused on the back end or operational aspects. I would say, that could be one area that to look at. And I think in supply chain, there are lots of things, such as, logistic planning. Damage detection that doesn't directly interact with the customer. So I think it just depends on how much of your data is of high quality and then narrow down your experiment to that.

Jimmy and James (2): 34:21

super. Thanks, Alex. Well, I think you two have competed for the quote of the day, which was either James's buzzword du jour or Alex's without high quality data. This is just fluff, thank you, Alex, for taking us through a quick canter through, generative AI and a little bit of education on what it actually is. some examples and some fantastic advice on how you can set yourselves up for success on this potentially, life changing new technology. usually at this point, we say to people, you know, if they want to know more, get in touch and we'll help them on this occasion, we will say, if you want to know more about AI, then get in touch with Alex and we'll put the contact details in the, in the blurb for the episode. because he's the expert. However, if you want to know about more about management fluff, we're quite good at that. We can do the management fluff bits,

Alex: 35:14

I think the three of us can do a great job together then.

Jimmy and James (2): 35:18

It was really interesting. Thanks very much, Alex.

Alex: 35:21

really enjoyed it. Thank you for having me. Great stuff. Thank you.

Jimmy and James (2): 35:24

Cheers. Thanks, everyone. Have a great week

James: 35:27

We cover a whole host of topics on this

Jimmy: 35:30

Purpose to corporate jargon.

James: 35:31

but always focused on one thing, getting the job done well.

Jimmy: 35:35

Easier said than done. So, If you've got Unhappy customers or employees Bosses or regulators Breathing down your neck

James: 35:44

of control and your costs are spiraling and that big IT transformation project that you've been promised just keeps failing to deliver. So

Jimmy: 35:54

Can help if you need to improve your performance, your team's performance, or your organization's. Get in touch at jimmy at jobdonewell. com

James: 36:04

Or james at jobdonewell. com.